Ring Opening in Benzocyclobutene Derivatives. I. The Reaction of

I. The Reaction of trans-1,2-Diiodobenzocyclobutene with Dimethylamine1. MELVIN W. HANNA, and S. W. FENTON. J. Org. Chem. , 1961, 26 (5), pp 1371–13...
1 downloads 0 Views 510KB Size
M.4Y

1961

1371

RING O P E N I N G I N BENZOCYCLOBUTBNE DERIVATIVES. I

presence of ferric chloride. A mixture of 13.2 g. (0.05mole) of o-xylylene bromide and 170 ml. of benzene was treated with 6.0 g. (0.037 mole) of ferric chloride in the same manner as above. The products isolated were as follows: 6.6 g. (79 mole %) of diphenylmethane, b.p. 133-135" (13 mm.), and 3.8 g. (30mole %) of o-dibenzylbenzene. The reaction of o-xylylene bromide with benzene in the

[CONTRIBUTION FROM

THE

SCHOOL O F

presence of stannic chloride. Similar treatment of 13.2 g. (0.05mole) of o-xylylene bromide and 80 ml. of benzene with 11.9 g. (0.1 mole) of stannic chloride gave 1.5 g. (18 mole %) of diphenylmethane, besides 5.2g. of o-xylylene bromide which was recovered unchanged. K Y ~ T OJAPAN ,

CHEMISTRY O F THE

UNIVERSITY O F MINNESOTA]

Ring Opening in Benzocyclobutene Derivatives. I. The Reaction of trans-l,2-Diiodobenzocyclobutenewith Dimethylamine' MELVIN W. HANNA2

AND

S. W. FENTON

Received August 8, 1960 Treatment of trans-1,2-diiodobenzocyclobutenewith dimethylamine leads to the formation of a,a,a'-trisdimethylaminoo-xylene and a small amount of a-dimethylamino-o-tolualdehyde.The formation of these products is rationalized by postulating an opening of the four-membered ring initiated by nucleophilic attack of dimethylamine.

Benzocyclobutene and its derivatives have lately become of interest for two reasons: First, these compounds are highly strained due to the presence of a fused benzene and four-membered ring system; second, benzocyclobutene derivatives are possible precursors of benaocyclobutadiene, a compound of great theoretical interest for which some molecular orbital calculations predict a substantial delocalization energy.3 Due to the strain inherent in the benzocyclobutene nucleus, the prediction has been made that compounds containing this structural feature might not be capable of existence, or a t best be highly reactive.* However, a whole series of benzocyclobutene derivatives, including benzocyclobutene itself, have been prepared and have been found to be quite stable.5+ Among these stable benzocyclobutene derivatives were the dihalobenzocyclobutenes, 1,2-dibromobenzocyclobutene (I) and 1,2-diiodobenzocyclobutene (11). These dihalides were stable towards boiling alcohol, alcoholic potassium acetate, and maleic anhy-

I. X 11. x

= =

Br

I

dride.5 Catalytic reduction did not rupture the four-membered ring but, instead, produced benzocyclobutene.' It was only when dibromide I was heated with bromine at 150" for twelve hours that ring opening occurred to . give a,a,d,a'-tetrabromo-o-~ylene.~~~ It has recently been found in this laboratory, however, that dihalides I and I1 undergo a facile ring opening when treated with secondary amines. This reaction was first observed when a mixture of dibromide I and diiodide I1 was treated with piperidine. Upon heating to 50" a mixture of piperidinium bromide and iodide precipitated and an undistillable red oil was obtained. In the subsequent investigation of this reaction, pure transdiiodide I1 was used to eliminate the possibility of different reactions from different stereoisomers and to make possible the determination of the (1) This paper is based on portions of a thesis submitted amount of amine salt formed. Dimethylamine was by M. W. Hanna to the Graduate School of the University of Minnesota in partial fulfillment of the requirements for employed as the amine in order to make the products readily distillable. the degree of Doctor of Philosophy. (2) National Science Foundation Predoctoral Fellow, When trans-diiodide I1 was treated with an 1958-59. Present address: California Institute of Tech- excess of dimethylamine in an Adkins bomb at 60" nology, Pasadena 4,Calif. for six hours, a mixture of two liquid products (3) J. D. Roberts, A. Streitweiser, Jr., and C. M. Regan, was formed along with two equivalents of dimethylJ . A m . Chem. Soc., 7 4 , 4579 (1952). (4) W.Baker (The Tilden Lecture, 1944),J . Chem. SOC., ammonium iodide. The lower boiling product, 258, (1945). formed in 10% yield, was shown to be a-dimethyl( 5 ) M. P. Cava and D. R. NaDier, J . Am. Chem. Soc., 79, amino-o-tolualdehyde (111); evidence for the struc1701 (1957). (6) M. P. Cava, R. L. Little, and D. R. Napier, J . A m . ture of this compound is presented below. The major product, formed in 64% yield, was a highly reChem. SOC.,80,2257(1958). ~I

(7) M. P. Cava and D. R. Napier, J . Am. Chem. SOC., (9) Some other examples of ring opening in benzocyclo80, 2255 (1958). ( 8 ) hl. P. Cava and I). R. Napier, J. A m . Chem. SOC., butene derivatives have recently been reported. See M. P. Cava and K. Muth, J . A m . Chem. SOC.,82, 652 (1960). 79, 3606 (1957).

1342

A "

AND FENTON

active liquid that began to decompose within fifteen minutes after distillation even though stored under argon. If this highly reactive liquid was reduced immediately after distillation using Raney nickel and dry Skellysolve F, a new amine was formed. This new amine was shown to be alabisdimethylamino-o-xylene (IV) by comparison of its properties with those of an authentic sample prepared by the method of von Braun and Cahn.Io The infrared spectrum of diamine IV prepared by this method was identical with that of the reduction product. A mixture melting point determination of the two dipicrates showed no depression. If the highly reactive liquid was treated with dilute acid and then with base, a product different from the starting material was obtained. An infrared spectrum and a mixture melting point determination of its picrate showed that this liquid product was identical with the lower boiling product obtained in the original reaction of diiodide I1 with dimethylamine; therefore, it was the aminoaldehyde 111. Assignment of structure I11 to the product resulting from acid treatment of the highly reactive liquid and to the lower boiling product from the original reaction was based on the following evidence : combustion analysis of the pure liquid was consistent with the formula CloHlaNO; both a picrate and a 2,4-dinitrophenylhydrazonederivative were formed, and the analyses of these derivatives were also consistent with products derived from a compound having the composition GOIII3NO. Three structures containing a benzene ring and a dimethylamino group attached to a side chain can be written with this empirical formula. These possibilities are wdimethylaminoacetophenone (V) and a-dimethylaminophenylacetaldehyde (VI).

YOL.

26

C o n h a t i o n that dimethylamine was lost in the conversion of the reactive liquid to aminoaldehyde I11 was obtained by treating the original reactive product with picric acid. After five r e crystallizations, analytically pure dimethylamine picrate was obtained whose melting point was not depressed upon admixture with an authentic aample. On the basis of these reactions there are two possible structures for the major product of the r e action of tramdiiodide I1 with dimethylamine. 42 Bisdimethylaminobenzocyclobutene (VII) would be expected on the basis of the structure of the starting material and that of the reduction product diamine 1V; a,a,a'-trisdimethylaminoo-xylene (VIII) would be reasonable on the basis of the aciddecomposition product, aminoaldehyde 111.

-

CHzN(CHa)z "(cH3)2 VI1

N (CHs)P

Q

L

H

(CH3)z]2

VI11

The correct structure for this product can be decided upon by a consideration of its proton magnetic resonance spectrum. The spectra of both diamine VI1 and triamine VI11 will contain absorptions due to the methyl and aromatic ring hydrogens, but these will appear in the high and the low field portions of the spectra respectively. These absorptions will. be neglected in the following discussion since they are common to both compounds. Diamine VII, then, should exhibit only one peak in its proton-resonance spectrumthat due to the two equivalent benzocyclobutenyl protons. Triamine VIII, however, should contain two peaks with approximately a 2 : l ratio of intensities, corresponding to the absorptions of the benzyl and benzylidine hydrogens respectively. The observed spectrum was exactly that expected for triamine VIII, having one absorption at 6.607 and one at 5.87s.11The absorption at 6.607 I11 IV had twice the intensity of the 5.877 peak and was, 9 N(CHa)z therefore, due to the two benzyl hydrogens, while O/CCHzN(CHa)z the absorption at 5.877 was due to the one bensylidine hydrogen. Additional evidence that this was V VI the proper assignment came from the observation that the benzyl protons benzyldimethylamine Amino aldehyde I11 reduced Tollens reagent have a resonance peak atin 6.687.12 Furthermore, readily, and its infrared spectrum exhibited a band in the proton resonance spectrum of bisdimethylat 1690 cm.-1, indicating the presence of an aro- aminomethane, the resonance peak due to the matic aldehyde. Furthermore, orth,odisubstitu- methylene protons was shifted to a lower value. tion would be expected in a product derived from of the magnetic field with respect to the methyl diiodide 11; and a strong band in the infrared spectrum at 760 cm.-l, coupled with the absence of a (11) For a definit,ion of 7 see G. V. D. Tiers, J . Phgs. strong band at 700 cm.-I, confirmed this. Only Chem., 62, 1151 (1958). structure 111 is consistent with all of the above (12) G. V. D. Tiers, Characteristic Nuclear Magnetic evidence, Resonance "Shielding" Values (Spectral Positions) for Hvdro-

CYCHCH0

(10) J. von Braun and R. 8. Cahn, Ann., 436, 262 ( 1924).

g e n i n Organic Structures, Central Research Department, Minnesota Mining and Manufacturing Co., St. Paul, Minn., March 28, 1968, p. 16.

MAY 1961

1373

RING OPENING IN BENZOCYCLOBUTENE DERIVATIVES. I

protons by a factor of 0.447." This chemical shift can be considered an approximate measure of the effect of adding a second dimethylamino group to a saturated carbon atom. It would be expected, then, that the absorption due to the benzylidine proton in triamine VI11 would be shifted to a lower field value by a factor of comparable magnitude. In agreement with this expectation a shift of 0.737 to a lower value of the applied field was observed. Thus NMR provides conclusive evidence for the assignment of structure VI11 to the highly reactive liquid obtained in the reaction of transdiiodide I1 with dimethylamine.'* The occurrence of a ring opening in the reaction of transdiiodide I1 with dimethylamine was supprising in view of the stability of its fowmembered ring toward other reagents. Under the conditions of its formation it must remain in the presence of a large excess of iodide ion a t 80" for eight to ten days, Yet, no ring opened product has been reported, and the yield of diiodide I1 is often as high as 90%. This indicates that diiodide 11, like dibromide I, is fairly resistant to ring opening by nucleophilic reagents. Dimethylamine is a more powerful nucleophilic reagent than iodide ion and, in addition, is a fairly strong base. Its action on transdiiodide I1 could be due therefore to either one or a combination of these properties. Since it is known that dibromide I w&s converted via an elimination-addition mechanism to 5-bromobenao[albiphenylene by treatment with potassium t butoxide,15 it is possible that a cis elimination of hydrogen iodide is the first step in the reaction of diiodide I1 with dimethylamine. To test the reactivity of this diiodide toward a base comparable with dimethylamine, a sample was heated in refluxing triethylamine for twenty-four hours. If elimination is important in this reaction, a compound similar to the above biphenylene derivative should be formed. At the completion of this reaction, however, there was no evidence of formation of triethylammonium iodide, and pure diiodide I1 was recovered in 85% yield. It is probable, then, that the nucleophilic character of dimethylamine was responsible for the ring opening in trans-diiodide 11. A possible reaction sequence is shown in Fig. 1. An attack of a dimethylamine molecule on diiodide IT, instead of displacing an iodine atom, produces ring opening resulting in the formation of carbanion IXa. Shift of a proton would then give crdimethylaminoa,a'diiodo-o-xylene (IXb) , and this intermediate would quickly react with the excess dimethylamine to give triamine VIII.

8

I

IXa

The initial step in this reaction sequence involves the breaking of a carbon-carbon bond and the formation of a carbon-nitrogen bond. A consideration of bond energies indicates that the initial product, carbanion IXa, would have a higher energy than the starting material by a factor of 14 kcal./ mole. However, considerable strain is relieved when the benaocyclobutene ring is broken, and this would act to lower the energy of intermediate IX. A rough estimate of the strain energy in the four-membered ring of benaocyclobutene derivatives, using the force constants for the in-plane C-H deformation in benzene and a similar deformation in methane, indicated a minimum strain energy of 20 kcal./mole. Thus, the energy gained by relieving the strain present in diiodide I1 is enough to make the first step in the above reaction sequence thermodynamically feasible. If the mechanism described above is correct, then ring opening should occur upon attack by a nucleophilic reagent whenever the bond being formed between the nucleophilic reagent and the carbon atom of the four-membered ring has an energy greater than 60-65 kcal./mole. This postulate is consistent with the observation that diiodide I1 is stable in the presence of a large excess of iodide ion since the carbon-iodine bond energy in benzyl iodide is only 43 kcal./mole. EXPERIMENTAL^^ tram-l,%Diiodobenaocyclobutene (11) waa prepared by the method of Coleman and Jensen,l' m.p. 62-64' (lit. m.p. 63-64').

.Treatmentof trana-l,d-diwdobenzocyclobutenewith dimethyk amine. Preparation of a-dimethylamino-o-toluald.ehy& (111) and a,u,a'-~risdimethylaminoo-zylene(VIII). trans-l,%Diiodobenaocyclobutene ( 17.8 g., 0.05 mole) and anhydrous dimethylamine (22.5 g., 0.5 mole) were placed in a glasslined Adkins bomb and heated a t 50-60' for 6 hr. The bomb w m allowed to cool overnight. The yellow oily product was transferred to a 500-ml. Erlenmeyer flaak with the aid of

(14) In a previous communication, S. W. Fenton and M. W. Hanns, J . Org. Chem., 24, 579 (1959), this compound waa erroneously reported to be diamine VII. (15) M. P. Cava and J. F. Stucker, J . Am. Chent. Soc.,

(16) Melting points were determined in a silicone oil bath and are uncorrected; infrared spectra were determined on a Perkin-Elmer Model 21 double-beam spectrophotometer; microanalyses were performed at the University of Minnesota by Mrs. Olga Hamerston and Mr. William C. K u~ N ~_ . . _ (